Researchers from the University Hospitals of Geneva and the University of Geneva have developed a groundbreaking crosslingual annotation projection methodology to address the shortage of high-quality annotated datasets in specialized fields like medicine. Utilizing a language-agnostic approach based on BERT technology, they effectively transferred annotations between languages. This led to the creation of FRASIMED, a vast French annotated resource containing over 2,000 synthetic clinical cases, integrated with datasets like CANTEMIST and DISTEMIST.
Read the full story — https://news.superagi.com/2023/09/20/researchers-unveil-crosslingual-annotation-methodology-boosting-clinical-french-nlp-resources/
Top comments (0)